Merge remote-tracking branch 'origin/master' into guoguo/linuxBuildFix

BrainScript/BrainScriptEvaluator.cpp

@@ -1164,7 +1164,7 @@ namespace Microsoft { namespace MSR { namespace BS {
                 // Like in an [] expression, we do not evaluate at this point, but pass in a lambda to compute on-demand.
                 let & args = argsExpr->args;
                 if (args.size() != lambda->GetNumParams())
-                    Fail(wstrprintf(L"function expects %d parameters, %d were provided", (int)lambda->GetNumParams(), (int)args.size()), argsExpr->location);
+                    Fail(wstrprintf(L"function expects %d positional parameters, %d were provided", (int)lambda->GetNumParams(), (int)args.size()), argsExpr->location);
                 vector<ConfigValuePtr> argVals(args.size());
                 //bool onlyOneArg = args.size() == 1 && argsExpr->namedArgs.empty();
                 for (size_t i = 0; i < args.size(); i++)    // positional arguments

CNTK.sln

@@ -195,14 +195,8 @@ Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Data", "Data", "{5F733BBA-F
 EndProject
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "LSTM", "LSTM", "{19EE975B-232D-49F0-94C7-6F1C6424FB53}"
 	ProjectSection(SolutionItems) = preProject
-		Tests\Speech\LSTM\baseline.cpu.txt = Tests\Speech\LSTM\baseline.cpu.txt
-		Tests\Speech\LSTM\baseline.gpu.txt = Tests\Speech\LSTM\baseline.gpu.txt
-		Tests\Speech\LSTM\baseline.windows.cpu.txt = Tests\Speech\LSTM\baseline.windows.cpu.txt
-		Tests\Speech\LSTM\baseline.windows.gpu.txt = Tests\Speech\LSTM\baseline.windows.gpu.txt
 		Tests\Speech\LSTM\cntk.config = Tests\Speech\LSTM\cntk.config
 		Tests\Speech\LSTM\lstmp-3layer_WithSelfStab.ndl = Tests\Speech\LSTM\lstmp-3layer_WithSelfStab.ndl
-		Tests\Speech\LSTM\run-test = Tests\Speech\LSTM\run-test
-		Tests\Speech\LSTM\testcases.yml = Tests\Speech\LSTM\testcases.yml
 	EndProjectSection
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "ParseConfig", "MachineLearning\ParseConfig\ParseConfig.vcxproj", "{7C4E77C9-6B17-4B02-82C1-DB62EEE2635B}"
@@ -334,6 +328,26 @@ Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Kaldi2Reader", "Kaldi2Reade
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "SequenceTraining", "MachineLearning\SequenceTraining\SequenceTraining.vcxproj", "{EAD17188-072C-4726-B840-A769C36DAD1B}"
 EndProject
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Truncated", "Truncated", "{88F85A64-105D-4CDA-8199-B7A312FC8A27}"
+	ProjectSection(SolutionItems) = preProject
+		Tests\Speech\LSTM\Truncated\baseline.cpu.txt = Tests\Speech\LSTM\Truncated\baseline.cpu.txt
+		Tests\Speech\LSTM\Truncated\baseline.gpu.txt = Tests\Speech\LSTM\Truncated\baseline.gpu.txt
+		Tests\Speech\LSTM\Truncated\baseline.windows.cpu.txt = Tests\Speech\LSTM\Truncated\baseline.windows.cpu.txt
+		Tests\Speech\LSTM\Truncated\baseline.windows.gpu.txt = Tests\Speech\LSTM\Truncated\baseline.windows.gpu.txt
+		Tests\Speech\LSTM\Truncated\run-test = Tests\Speech\LSTM\Truncated\run-test
+		Tests\Speech\LSTM\Truncated\testcases.yml = Tests\Speech\LSTM\Truncated\testcases.yml
+	EndProjectSection
+EndProject
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "FullUtterance", "FullUtterance", "{8241108A-7824-4FF2-BECA-7521A9D89DCF}"
+	ProjectSection(SolutionItems) = preProject
+		Tests\Speech\LSTM\FullUtterance\baseline.cpu.txt = Tests\Speech\LSTM\FullUtterance\baseline.cpu.txt
+		Tests\Speech\LSTM\FullUtterance\baseline.gpu.txt = Tests\Speech\LSTM\FullUtterance\baseline.gpu.txt
+		Tests\Speech\LSTM\FullUtterance\baseline.windows.cpu.txt = Tests\Speech\LSTM\FullUtterance\baseline.windows.cpu.txt
+		Tests\Speech\LSTM\FullUtterance\baseline.windows.gpu.txt = Tests\Speech\LSTM\FullUtterance\baseline.windows.gpu.txt
+		Tests\Speech\LSTM\FullUtterance\run-test = Tests\Speech\LSTM\FullUtterance\run-test
+		Tests\Speech\LSTM\FullUtterance\testcases.yml = Tests\Speech\LSTM\FullUtterance\testcases.yml
+	EndProjectSection
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Win32 = Debug|Win32
@@ -503,5 +517,7 @@ Global
 		{3E9C89B1-C045-4F42-92B2-F9FFFFC2DBD4} = {39E42C4B-A078-4CA4-9D92-B883D8129601}
 		{C70E1572-20FF-496C-A0A9-10AA6755A07C} = {39E42C4B-A078-4CA4-9D92-B883D8129601}
 		{EAD17188-072C-4726-B840-A769C36DAD1B} = {DD043083-71A4-409A-AA91-F9C548DCF7EC}
+		{88F85A64-105D-4CDA-8199-B7A312FC8A27} = {19EE975B-232D-49F0-94C7-6F1C6424FB53}
+		{8241108A-7824-4FF2-BECA-7521A9D89DCF} = {19EE975B-232D-49F0-94C7-6F1C6424FB53}
 	EndGlobalSection
 EndGlobal

DataReader/UCIFastReader/UCIFastReader.cpp

@@ -848,7 +848,10 @@ bool UCIFastReader<ElemType>::GetMinibatch(std::map<std::wstring, Matrix<ElemTyp
                 }
             }
         }
-    } 
+    }
+
+    // create the respective MBLayout
+    m_pMBLayout->Init(1, actualmbsize, false/*means it is not sequential*/);
 
     // if we are writing out to the caching writer, do it now
     if (m_cachingWriter)

DataReader/UCIFastReader/UCIFastReader.h

@@ -69,11 +69,12 @@ class UCIFastReader : public IDataReader<ElemType>
 
     bool m_endReached;
     int m_traceLevel;
-   
+
     // feature and label data are parallel arrays
     std::vector<ElemType> m_featureData;
     std::vector<LabelIdType> m_labelIdData;
     std::vector<LabelType> m_labelData;
+    MBLayoutPtr m_pMBLayout;
 
     // map is from ElemType to LabelType
     // For UCI, we really only need an int for label data, but we have to transmit in Matrix, so use ElemType instead
@@ -106,13 +107,13 @@ class UCIFastReader : public IDataReader<ElemType>
 public:
     virtual void Init(const ConfigParameters& config);
     virtual void Destroy();
-    UCIFastReader() { m_featuresBuffer=NULL; m_labelsBuffer=NULL; m_labelsIdBuffer=NULL; }
+    UCIFastReader() { m_featuresBuffer=NULL; m_labelsBuffer=NULL; m_labelsIdBuffer=NULL; m_pMBLayout=make_shared<MBLayout>(); }
     virtual ~UCIFastReader();
     virtual void StartMinibatchLoop(size_t mbSize, size_t epoch, size_t requestedEpochSamples=requestDataSize);
     virtual bool GetMinibatch(std::map<std::wstring, Matrix<ElemType>*>& matrices);
 
     size_t GetNumParallelSequences() { return mBlgSize; }
-    void CopyMBLayoutTo(MBLayoutPtr){};
+    void CopyMBLayoutTo(MBLayoutPtr pMBLayout) { pMBLayout->CopyFrom(m_pMBLayout); };
     virtual const std::map<LabelIdType, LabelType>& GetLabelMapping(const std::wstring& sectionName);
     virtual void SetLabelMapping(const std::wstring& sectionName, const std::map<LabelIdType, LabelType>& labelMapping);
     virtual bool GetData(const std::wstring& sectionName, size_t numRecords, void* data, size_t& dataBufferSize, size_t recordStart=0);

MachineLearning/CNTK/NDLUtil.h

@@ -101,6 +101,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                 // make sure to clear the caches so we pick up the new nodes
                 m_net->ClearCaches();
                 // validate the network
+                //m_net->SetFakeMBLayoutForValidation();  // fake an MB layout to match the initial values of Input
                 if (dumpFileName != L"")
                     m_net->DumpAllNodesToFile(false, dumpFileName, false);
                 m_net->ValidateNetwork(!fullValidate);

MachineLearning/CNTK/SimpleNetworkBuilder.cpp

@@ -1739,7 +1739,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                     input = output;
                 }
 
-                size_t idim = input->FunctionValues().GetNumRows();
+                size_t idim = input->GetNumRows();
                 assert(m_lookupTabelOrderSizes.size() == m_streamSizes.size());
 
                 e = builder.CreateLearnableParameter(msra::strfun::wstrprintf(L"Embedding%d", idx), m_layerSizes[1], idim / m_lookupTabelOrderSizes[idx]);
@@ -2060,7 +2060,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                     input = output;
                 }
 
-                size_t idim = input->FunctionValues().GetNumRows();
+                size_t idim = input->GetNumRows();
                 assert(m_lookupTabelOrderSizes.size() == m_streamSizes.size());
 
                 e = builder.CreateLearnableParameter(msra::strfun::wstrprintf(L"Embedding%d", idx), m_layerSizes[1], idim / m_lookupTabelOrderSizes[idx]);
@@ -2433,7 +2433,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
             else
                 std::runtime_error("Output layer size must be specified when converting pretrained network, use outputLayerSize=");
 
-            size_t penultimateSize = input->FunctionValues().GetNumRows();
+            size_t penultimateSize = input->GetNumRows();
 
             wstring nameOfW = msra::strfun::wstrprintf(L"W%d", i);
             wstring nameOfB = msra::strfun::wstrprintf(L"B%d", i);